Aims A review of practice to determine whether serum-free light chain (SFLC) assays are helpful in detecting underlying clonal B-cell disorders or amyloidosis in patients with primary antibody deficiency (PAD) and recurrent infection.
Methods SFLC were assayed by nephelometry (BN2 nephelometer, Siemens; FREELITE assay, Binding Site). We reviewed SFLC test results recorded in our regional laboratory over a 4-year time period; 20 adults with PAD were identified as having been tested on at least two occasions.
Results Of 20 patients, 4 with PAD had abnormal serum-free kappa/lambda (K/L) ratios but no evidence of B-cell clonality. We also found extremely low levels of kappa and or lambda (below the limits of reliable detection) in 19/20 PAD cases (mostly common variable immunodeficiency), such that in many, ratios were not calculable.
Conclusions The data suggest that the abnormal ratios are generated by an inability to produce and/or secrete SFLCs, particularly kappa FLC. In this small initial study, we seek to highlight PAD cases where a suspicious K/L ratio, typically with very low absolute quantities of SFLCs, most likely points to B-cell dysfunction, rather than to B lymphocyte clonality.
- Laboratory Tests
Statistics from Altmetric.com
Commercial serum-free light chain (SFLC) assays are widely used in routine pathology laboratories. They are useful both in the investigation of possible B lymphocyte clonal disorders,1 ,2 and in cases of immunoglobulin light chain (AL) amyloid, where the SFLC assay has been shown to be a more sensitive diagnostic tool than conventional serum electrophoresis and serum immunofixation.3
In normal individuals, small amounts of both free lambda and free kappa light chains are found and detected by SFLC assays. The assay employs polyclonal antibodies specific for cryptic epitopes on immunoglobulin light chains normally concealed within the tertiary structure of an immunoglobulin molecule. The normal ratio of kappa/lambda (K/L) in an adult population is between 0.26 and 1.65.3 Values lying outside this range raise suspicion of a clonal B lymphocyte disorder.
Patients with primary antibody deficiencies (PAD) have an increased likelihood of developing malignancy, including non-Hodgkin's lymphomas.4 ,5 To confidently make a diagnosis of PAD, haematological malignancy with secondary immunoglobulin deficiency must be excluded.
If we exclude selective IgA deficiency (seen in around 1 in 500 persons), PAD is a rare disorder. Common variable immunodeficiency (CVID) is suspected in patients with recurrent bacterial infection and associated hypogammaglobulinaemia, and is a diagnosis of exclusion made once more common, secondary hypogammaglobulinaemia (drug related, nephritic syndrome, other) and B lymphocyte clonality with associated immune paresis) have been excluded.
The SFLC assay is being increasingly used in the diagnosis of B-cell neoplastic disorders. We report our analysis of results in 20 patients with PAD.
In a review of SFLC test results recorded in our regional laboratory over a 4-year time period, 20 adults with PAD were identified as having been tested on at least two occasions. They included 13 males, and were aged 18–65 years at diagnosis. All samples had been sent to the laboratory as part of routine clinical practice. Four patients were tested before starting pooled normal human immunoglobulin infusions. In two cases, treatment was via the subcutaneous route (15% product), whereas for the others it was intravenous, using 0.4 gm/kg body weight monthly. All patients were established on regular IgG infusions as infection prophylaxis. In patients on immunoglobulin replacement therapy, samples were taken immediately prior to their next routine infusion (in most cases 3 weeks after the previous infusion). In some cases, the laboratory requested the SFLC test because other routine tests (eg, low gamma region on serum electrophoresis) had raised laboratory concerns about possible B-cell malignancy. In this group of 20 cases, common variable immunodeficiency (CVID) was the commonest diagnosis (16/20). Two of the remaining cases (brothers) had X-linked agammaglobulinaemia (XLA). One male had Hyper IgM Syndrome (HIGM). The remaining female patient (aged 53 years) had late-onset antibody immunodeficiency syndrome with associated very high levels of polyclonal IgM where genetic testing had excluded classical HIGM. The common denominator for all 20 cases was antibody immunodeficiency and associated increased frequency of bacterial infection. Patient clinic follow-up ranged from 9 months to 18 years (PAD diagnosis preceding SFLC testing in most cases). No clinical or laboratory evidence suggesting B lymphocyte clonality has been found in any of the 20 cases to date. In all 20 cases, the abnormal SFLC results were checked and confirmed in at least two separate blood samples. In 4/20 cases, paired blood samples were obtained both before and several months after starting pooled normal immunoglobulin infection prophylaxis. In the other 16 cases, paired blood samples were taken up to 15 years after initial diagnosis/commencement on immunoglobulin infusions. The B lymphocyte counts quoted relate to the counts noted at the time of the first paired sample. The time interval between paired samples ranged between 3 months and 4 years. SFLC values were also determined in three different commercial preparations of pooled normal immunoglobulin prepared for intravenous or subcutaneous use.
SFLCs were quantified by nephelometry (FREELITE, The Binding Site Group Ltd, Birmingham, UK), according to the manufacture's instructions, using a BN2 nephelometer (Siemens Healthcare Diagnostics, Camberley, Surrey, UK). Between-batch variation was 7.1% at mean 17.0 mg/l for kappa SFLCs and 7.0% at 27.4 mg/l for lambda SFLCs.
B-cells were enumerated by standard flow cytometry using whole blood to quantify CD19 cells. Allophycocyanin-conjugated Anti-CD19 was supplied by Becton Dickinson, Cowley UK. Analysis was performed using a FacsCanto II flow cytometer (Becton Dickinson).
Pooled normal immunoglobulin from four manufacturers was prepared according to the manufacturer's instructions and analysed nephelometrically as per serum for immunoglobulin and SFLC content. These were: SUBCUVIA (Baxter Healthcare Ltd, Newbury, Berkshire, UK); Kiovig (Baxter Healthcare Ltd); Privigen (CLS Behring UK Ltd, Haywards Heath, West Sussex, UK); Flebogamma (Grifols UK Ltd, Cambridge, UK).
The commonest abnormal finding seen in 19/20 cases was of very low (in many cases undetectable) absolute levels for either lambda or kappa light chains (and in 15/20 cases low levels for both light chains). Values were either below the lower limit of detection, and/or significantly below the normal reference ranges for these assays (lambda 5.7–26.3 mg/l, kappa 3.3–19.4 mg/l) (figure 1A,B). In figure 1A, (kappa values) in one case (classical HIGM case), SFLC values were within the normal range on both occasions tested, and in one other case (non-HIGM with raised polyclonal IgM), the value was normal on one occasion. The other 18 cases had markedly low Kappa values on both occasions tested. Kappa FLC were always low in all the cases of CVID and the two cases of XLA.
In 18/20 cases, pretreatment immunoglobulins were available, the other two cases having transferred to our care after commencing immunoglobulin at other centres (records unavailable). Sixteen of 18 had a definitive IgG quantified; the remaining two were <0.5 g/l and <0.7 g/l. The median IgG of the 18 patients was 1.86 g/l (range <0.5–5.1 g/l). IgA was always low (range <0.05–0.46 g/l). IgM was raised in classical HIGM (IgG 1.82 g/l; IgM 3.6 g/l) and in non-HIGM with polyclonal raised IgM (IgG 2.77 g/l; IgA 0.46 g/l; IgM 5.02 g/l). IgM was normal in two other cases (IgG 1.70 g/l; IgM 0.95 and IgG <0.5 g/l; IgM 0.68 g/l), and low in the remaining 14 cases (range <0.05–0.41 g/l).
Figure 1B shows that lambda levels were more likely to be normal (seven cases) than kappa levels. These cases included the female with adult-onset hypogammaglobulinaemia and elevated polyclonal IgM, three CVID cases, one XLA case and the classical HIGM case. The male with classical HIGM (both kappa and lambda normal) was the only PAD case who tested normal on both occasions, as predicted (see discussion). With few exceptions, results for both lambda and kappa were stable over time (up to 4 years between paired samples).
Figure 2 shows that, while B lymphocyte counts varied from normal to below the reference range (marked), the circulating number of B lymphocytes showed no correlation with either SFLC concentration.
In figure 3, we show schematically the results for our 20 cases (at first time point), in the context of published normal and abnormal (B-cell clonal disorders), using a log scale, to contextualise and demonstrate that most PAD cases have strikingly low SFLC values. Four PAD cases show abnormal ratios with low kappa but normal lambda SFLCs. These fall outside of the typical areas for B-cell malignancy patients. In many, the ratio was incalculable owing to the limitations of assay sensitivity.
Pharmaceutical grade normal pooled IgG preparations were shown to contain detectable FLCs. In all cases, these were roughly proportionate to the IgG in the preparation (table 1), but all were relatively lower than seen in normal serum. Of note, our two patients on subcutaneous immunoglobulins therapy were at the severe end of the spectrum before commencing treatment (IgG 1.25 g/l and IgG 1.80 g/l). Both have a diagnosis of CVID. Both SFLCs were below the level of detection in the first patient on both occasions tested. In the second patient, kappa SFLCs were below the level of detection on both occasions; lambda was 8.9 mg/l and 8.2 mg/l. SFLC measurements preimmunoglobulin treatment are not available for these two cases.
An abnormal SFLC K/L ratio typically points to the presence of a B-cell clone.8 Indeed, some authors have argued that an abnormal ratio will ‘only be present in the context of a plasma cell dyscrasia or other B-cell lymphoproliferative disorder’.9 In contrast, Beetham et al found 7/932 patients with ‘false positive’ values.10 With the very low concentrations seen in this study, further caution in interpretation of an abnormal ratio is required. A small bias in either analyte may introduce a skew with a profound effect on the K/L ratio.
PAD cases are known to be at increased risk of developing B-cell clonal disease.4 ,5 By analogy with AL amyloid (eg, Katzmann et al3), we introduced use of the K/L ratio as a potential early marker of B-cell clonal disease in PAD. Our data suggest that this approach may not be helpful in the context of PAD.
The majority of our 20 cases had been treated with immunoglobulin replacement therapy before serum testing. Commercial preparations of pooled normal IgG did contain detectable kappa and lambda SFLCs, but it seems unlikely that they affect any of our conclusions. All our cases have normal renal function, and the half-life of FLC in the circulation is only 2–6 h. Therapeutic IgG infusions are typically repeated every 2–4 weeks, so that concentrations measured preinfusion are likely to only contain FLCs produced by the patient's immune system. Also, all cases tested prior to commencing immunoglobulin replacement therapy did not increase when repeated weeks/months after treatment had been initiated, and results were similar to the group in whom treatment had already been established. While it may be hypothesized that the absorption characteristics of FLCs, and hence, half-life in serum, might be different when replacement immunoglobulin is by the subcutaneous route, we saw no evidence of this. Both our patients showed very low kappa SFLCs, in keeping with the other CVID patients in this study.
In our Bristol-based laboratory, during this review of SFLC assay results, we noticed that PAD cases, with no evidence of malignancy, tended to have very low (often undetectable) SFLCs (especially kappa) with either abnormal K/L ratios or ratios which were difficult to calculate. Most of the PAD cases picked up in the review had known CVID. Low SFLC levels were unrelated to the peripheral B-cell count. We assume that in these cases with PAD and proven antibody deficiency, low SFLC reflects an intrinsic inability to manufacture or secrete immunoglobulin light chain. Some support for this might be drawn from the pretreatment immunoglobulin concentrations. IgG and IgA were low in all cases where pretreatment results were available (18 of 20 cases). Certainly the two cases with hyper-IgM had normal lambda FLCs, though kappa expression was variable. However, the other three cases seen with normal lambda SFLCs (figure 3) all had very low IgG, IgA and IgM at presentation (IgG 1.80, 2.85 and 3.70 g/l; IgA all <0.25 g/l; IgM all <0.21 g/l), so there is not a simple correlation.
Anecdotally, we see some patients treated with myeloablative therapy, in which the SFLCs may become undetectable, and this may be followed by a transient hypogammaglobulinaemia, suggesting again that lack of production of whole immunoglobulins is accompanied by a decrease in SFLCs. We have not formally reviewed these data, but it may warrant further study.
Hypogammaglobulinaemia is seen in perhaps a third of patients with chronic lymphocytic leukaemia (CLL) at some time in their disease. It is interesting to note that patients with a similar profile to our PAD patients (either both SFLCs low or low kappa with normal lambda SFLCs) may be seen in a minority (<0.5%) of patients with CLL.11 In this study, Pratt et al noted 39% of CLL patients to have an abnormal ratio, and that this was indicative of a subset of patients with poorer prognosis. By contrast with our PAD patients, in the vast majority of CLL patients with an abnormal SFLC ratio, this was due to an elevation of one of the light chains. No results for serum immunoglobulins were presented. Similarly, Tsai et al12 showed in a prospective study that abnormal FLC ratios were found in 38% of 109 patients before a diagnosis of CLL. Hypogammaglobulinaemia predated a diagnosis of CLL in 3%. No patients were identified in the study with FLC profiles as seen in our PAD patients. Further work is required to determine if the very low kappa SFLCs we see in the PAD patients also correlate with hypogammaglobulinaemia in CLL.
This report highlights the need for caution before presuming that a suspicious K/L ratio, with low absolute quantities of SFLCs, especially kappa, is due to B lymphocyte clonality. The SFLC test may provide a simple inexpensive tool to help deduce the definitive diagnosis of PAD. Further studies are required to determine whether results will reliably be normal in HIGM (where the defect is one of failure to isotype switch from IgM to IgG), while abnormal in XLA, CVID, and certain other antibody deficiency syndromes.
Until now, an abnormal ratio in the serum FLCs (SFLC) has been almost exclusively described in the context of B-cell malignancy, and is widely regarded as a marker of clonal B lymphocyte disorders. We show here that this must be viewed with caution in the context of PAD; a group of immunodeficient patients known to have a high prevalence of B-cell malignancy, but where SFLC is an unreliable test for clonality. Also, an incidental finding of very low SFLCs should raise suspicion of PAD, and this test may prove to be useful in identifying cases of suspected or evolving antibody immunodeficiency, though this requires further study.
Abnormal kappa/lambda ratios, as detected by serum-free light chain (SFLC) assays, are widely used to detect B-cell malignancies.
Abnormal kappa/lambda ratios are commonly seen in primary immunodeficiency, associated with underproduction of SFLCs.
Abnormal ratios seen in the context of PAD should not be considered as evidence of B-cell malignancy.
Our thanks go to Ellen Roberts for her technical expertise, and to Lisa Smith for help with patient samples and infusion products.
Contributors DJU: designed the study; RJL: analysed the data; MJW: performed analyses for serum-free light chains. All authors contributed to the writing of the paper and approved the final draft.
Funding We thank the North Bristol NHS Trust Showering Fund (SF 80) for supporting the early part of this study.
Competing interests None.
Ethics approval This study was undertaken as a retrospective review of practice of tests performed during the routine workup of these patients. Consent for analysis was given in clinic. No patients are identifiable in this paper.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.